A bayesian approach to bioassay. technical report m498
作者:
Richard M. Fagerstrom,
Duane A. Meeter,
期刊:
Communications in Statistics - Simulation and Computation
(Taylor Available online 1983)
卷期:
Volume 12,
issue 5
页码: 541-558
ISSN:0361-0918
年代: 1983
DOI:10.1080/03610918308812341
出版商: Marcel Dekker, Inc.
关键词: bioassay;multiple regrssion;average level estimator;interpolated estimator;Bayesian inference;fractional factorial;variance reduction techniques;efficiency
数据来源: Taylor
摘要:
We describe the design and analysis for a simulation experiment to compare the mean-squared errors (MSE's) of two quantile estimators defined for random walk designs. The dependence of the easily computed MSE of the first estimator on the levels of five factors is examined via multiple regression. This information is used to plan a simulation to compute the MSE of the second estimator using a fraction of a 3352factorial allowing uncorrelated estimates for all main effects and the two-factor interactions of a specified factor. Efficient estimation of the MSE of the second estimator is attempted through antithetic and control variate techniques of variance reduction, with modest success.
点击下载:
PDF (546KB)
返 回